Large volume ecg sensor data classification and association rules

RESULT
Pretreat
The aim of this study is to develop a machine learning model for classifying electrocardiogram (ECG) signals into five categories: Normal, Unknown, Ventricular Ectopic Beat, Supraventricular Ectopic Beat, and Fusion Beat while considering association rules. The model includes functions for adding Gaussian noise to the ECG signals, defining and training a convolutional neural network (CNN) model on the ECG data, and evaluating the performance of the trained model using metrics such as accuracy and confusion matrix.
The CNN model architecture includes several convolutional layers, max pooling layers, and fully connected layers. The model is trained on the ECG data using the categorical cross-entropy loss and the Adam optimizer. The performance of the trained model is evaluated using accuracy and visualizations of the training and validation loss and accuracy over time Figure [4]. The confusion matrix is also displayed to provide a detailed breakdown of the model's performance across each category. The Accuracy of this model is 98.09%.



Figure 4. Accuracy and Loss models

The next method defines a function that takes a confusion matrix and class labels as input and plots a visualization of the matrix using matplotlib. The confusion matrix is computed using the function from scikit-learn library, which takes the true labels and predicted labels as input Figure [5].

Figure 5. Confusion matrix, with normalization

The function has several optional parameters, including normalize to normalize the confusion matrix, title to set the title of the plot, and set the color map of the plot. The function uses itertools to iterate over the rows and columns of the confusion matrix and plot the values in each cell. If normalize is set to True, the function normalizes the values in the confusion matrix by dividing each row by its sum. Finally, the method uses plt methods to create a color-coded visualization of the confusion matrix with labeled axes and a color bar. The class labels are also displayed on the x and y axes.